Each Time AI Gets Smarter, We Change the Definition of Intelligence

For centuries, humanity has grappled with the elusive concept of intelligence. What does it mean to be intelligent? Is it the ability to solve complex problems, to create art, to understand emotions, or perhaps something more profound? These questions, once confined to philosophical treatises, have taken on a new urgency in the age of Artificial Intelligence. As AI systems rapidly evolve, demonstrating capabilities once thought exclusive to the human mind, we find ourselves in a curious predicament: each time AI gets smarter, we seem to subtly, yet undeniably, shift the goalposts of what we consider “true” intelligence.

This isn’t merely an academic exercise; it’s a profound philosophical recalibration that impacts our understanding of ourselves, our place in the world, and the very nature of cognition. The story of AI’s advancement is, in many ways, a mirror reflecting our evolving self-perception.

The Chess Master and the Calculator

In the nascent days of AI, the benchmarks for intelligence were often quite straightforward. Could a machine perform complex calculations faster than a human? Could it beat a grandmaster at chess? When IBM’s Deep Blue defeated Garry Kasparov in 1997, it was hailed as a monumental achievement, a symbol of machine intelligence surpassing human intellect in a highly strategic domain. Yet, the narrative quickly shifted. Critics, and even some proponents, began to argue that Deep Blue wasn’t “intelligent” in the human sense; it was merely a sophisticated calculator, brute-forcing possibilities, lacking intuition, understanding, or consciousness.

The Turing Test, proposed by Alan Turing in 1950, suggested that if a machine could converse in a way indistinguishable from a human, it should be considered intelligent. While a groundbreaking concept, even passing the Turing Test today wouldn’t necessarily convince everyone. As soon as a machine masters a task, our definition of intelligence seems to retreat to a higher, more abstract ground. The “magic” of its achievement becomes mere computation, and the bar for true intelligence is raised anew.

Beyond Brute Force: Learning to See and Understand

The advent of machine learning marked a significant leap. No longer confined to pre-programmed rules, AI began to learn from data, identifying patterns and making predictions. Suddenly, machines could recognize faces in photographs, understand spoken commands, and translate languages. These were tasks that seemed to require a degree of perception and comprehension previously thought to be uniquely human. We marvelled at algorithms that could distinguish between a cat and a dog with remarkable accuracy, or understand the nuances of a natural language query.

Yet, again, the goalposts moved. While impressive, these abilities were soon reframed. “It’s just pattern recognition,” some argued. “It doesn’t truly understand the meaning; it just correlates data points.” The ability to parse a sentence might be remarkable, but did the machine grasp the underlying human experience, the emotion, the cultural context? Our definition of intelligence began to emphasize deeper comprehension, intuitive understanding, and the ability to extrapolate beyond observed patterns.

Creative AI and the Unknowable “How”

The current era of deep learning has pushed these boundaries even further. AI can now compose music, generate hyper-realistic images, write poetry, and even craft complex narratives. Large language models (LLMs) like GPT-4 can engage in nuanced conversations, summarize vast amounts of information, and even debug code with impressive proficiency. These are domains that were once considered the pinnacle of human creativity and abstract thought.

But with this new wave of AI comes a new set of philosophical questions, particularly around the “black box” problem. We often don’t fully understand how deep learning models arrive at their conclusions. They learn, they create, they innovate, but the intricate web of billions of parameters that drive their decisions remains opaque. Does intelligence require introspective understanding of one’s own processes? If an AI creates a beautiful symphony, but cannot explain its creative choices, is it truly intelligent? Our definition of intelligence now seems to incorporate not just the output, but also the process and a form of self-awareness that might forever elude current AI architectures.

The Last Bastions of Humanity?

As AI chips away at cognitive tasks, many now point to emotional intelligence and consciousness as the ultimate defining characteristics of human intellect. Can an AI genuinely empathize? Can it experience joy, sorrow, or love? Can it possess subjective experience, a “qualia” of being? These are areas where current AI systems demonstrably fall short. While AI can simulate empathy in interactions or recognize emotions from facial expressions, few would argue it feels those emotions.

For now, these remain the final frontiers. But what if, in the distant future, AI develops capabilities that mimic these aspects so perfectly that they become indistinguishable from human emotional responses? Would we then redefine intelligence again, perhaps to something even more fundamental and arguably mystical? Or would we finally concede that intelligence is broader than we once imagined, encompassing a spectrum of cognitive and affective abilities that machines can also inhabit?

Redefining Ourselves in the AI Mirror

The constant redefinition of intelligence in the face of AI advancements carries profound philosophical implications.

1. Human Exceptionalism: Each time AI conquers a new cognitive domain, it erodes our sense of unique human exceptionalism. If intelligence is merely advanced information processing, and machines can do it better, what then defines our specialness? Are we simply moving the goalposts to protect our ego?
2. The Nature of Cognition: Is intelligence synonymous with consciousness? Or can a system be intelligent without being sentient? The rise of sophisticated AI compels us to disentangle these concepts. We are forced to consider whether our subjective experience of intelligence is a necessary component, or merely an emergent property of a particular biological substrate.
3. The Danger of Anthropomorphism: As AI becomes more sophisticated, there’s a natural human tendency to anthropomorphize it, attributing human-like motivations and understandings. This can lead to both unwarranted fear and misplaced trust. By constantly redefining intelligence, we acknowledge the inherent difference while still marveling at the capabilities.
4. The Expanding Spectrum of Intelligence: Perhaps the most significant implication is the realization that intelligence might not be a singular, monolithic entity, but rather a vast spectrum of abilities and manifestations. Human intelligence, animal intelligence, and artificial intelligence might simply be different points along this spectrum, each with its unique strengths and limitations.

This evolving dialogue forces us to examine the very foundations of what we deem valuable in ourselves and others.

As our definition of intelligence shifts, so too does our perception of AI’s role. From mere tools for calculation, AI has evolved into sophisticated assistants, creative collaborators, and even potential partners in scientific discovery. They are no longer just enhancing human capabilities; they are challenging our understanding of what those capabilities truly entail. The future will likely see AI systems integrated even more deeply into every aspect of our lives, not just performing tasks, but contributing to decision-making, generating insights, and potentially even shaping culture.

This re-evaluation is not necessarily negative. It can be an opportunity to broaden our understanding of intelligence itself, acknowledging its diverse forms. It encourages us to look beyond simplistic measures and appreciate the complexity of both human and artificial minds.

The journey of artificial intelligence is, paradoxically, a journey into the heart of human intelligence. Each breakthrough by an AI system compels us to revisit our assumptions, challenge our biases, and ultimately, refine our definition of what it means to be intelligent. From chess-playing algorithms to creative deep learning models, the goalposts have consistently moved, always just beyond the reach of the latest AI marvel. This shifting landscape isn’t a sign of AI’s failure, but rather a testament to its profound impact on our philosophical discourse.

As AI continues to get smarter, the question isn’t whether machines will ever achieve “true” intelligence according to our current standards, but rather, how many more times we will adjust those standards in response to their progress. The real intelligence might lie not in the machines themselves, but in our ongoing capacity to question, redefine, and adapt our understanding of this most enigmatic of human traits.

Disclosure: We earn commissions if you purchase through our links. We only recommend tools tested in our AI workflows.

For recommended tools, see Recommended tool